Not Applicable
Not Applicable
This invention relates to the field of amplifiers, and more particularly, to a high-fidelity audio amplifier having an improved floating bridge circuit topology.
It is well known among skilled high-fidelity audio amplifier designers that traditional performance measurements such as frequency response, total harmonic distortion, and output damping factor do not fully predict amplifier sound quality, particularly where high measured performance is achieved through the use of large amounts of global negative feedback. Some designers seeking to advance the art have therefore shifted their attention toward the intrinsic behavior of the amplifier prior to the application of feedback, believing that such xe2x80x9copen loopxe2x80x9d behavior better represents the performance of the amplifier once it is removed from the test bench, placed in a sound system, and judged by a listener. This has led to increased interest in amplifier topologies with the potential for high performance using little or no global negative feedback.
One such topology is the floating bridge or xe2x80x9ccirclotron,xe2x80x9d a unity gain current amplifier having a pair of gain devices and a pair of floating DC power supplies arranged in a balanced-bridge configuration. This current amplifier is typically combined with a preceding voltage amplifier to form a complete line-level or power amplification system. Audio amplifiers of this general type are well known in the prior art, the earliest examples being vacuum tube power amplifier circuits. U.S. Pat. No. 2,705,265 (1955) to Hall and U.S. Pat. No. 2,828,369 (1958) to Wiggins describe examples of such designs. More recently, solid state amplifiers using a floating bridge output stage have appeared. U.S. Pat. No. 4,229,706 (1980) to Bongiomo and U.S. Pat. No. 4,719,431 (1988) to Karsten describe floating bridge amplifiers using bipolar junction transistor (BJT) output stages, while U.S. Pat. No. 6,242,977 (2001) to Karsten describes a floating bridge amplifier using metal oxide semiconductor field effect transistor (MOSFET) gain devices.
The replacement of vacuum tubes with solid-state devices significantly improves the open-loop performance of a floating bridge output stage. The higher transconductance of solid-state devices results in more accurate tracking of the input signal voltage, while their superior aging characteristics allow for more reliable selection of matched gain devices for enhanced symmetry and more complete cancellation of evenorder distortion products. The solid-state floating bridge thus offers a firm foundation for linear open-loop design, particularly when the output stage operates in class A up to at least one-tenth of its maximum output power (the range where most listening occurs), or where practical, over its entire output power envelope.
In such an implementation, the major remaining performance limitation in a solid-state floating bridge amplifier is the nonlinear transfer of signal voltage from the preceding voltage amplifier output to the current amplifier input. In a BJT output stage, this nonlinearity results mainly from dynamic fluctuations in the current gain (beta) of the output devices, while in a power MOSFET output stage, it results mainly from dynamic fluctuations in the input capacitance of the output MOSFETs. Either condition is reflected as a nonlinear input impedance to the output stage that tends to distort the applied input signal.
An obvious solution to this problem which does not involve global negative feedback is to add a unity-gain emitter or source follower driver stage ahead of the floating bridge output stage. This reduces overall nonlinearity by essentially substituting the smaller nonlinearities of the driver gain devices for the larger nonlinearities of the output gain devices. However, the effectiveness of this approach has been limited in the past because the driver stage works under conditions not optimized for linear operation, and thus contributes significant distortion of its own. All driver circuits available under the prior art for use in a solid-state floating bridge share this limitation to some degree, and may introduce additional performance limitations.
Depending on configuration, these prior art driver circuits suffer from one or more of the following drawbacks:
a) Distortion contributed by the driver stage itself. In U.S. Pat. No. 4,107,619 (1978), Pass teaches that distortion in any amplifying device arises when the gain of the device is altered by changing voltage across the device and changing current through the device; and conversely, that distortion may be improved by reducing or eliminating such voltage and current changes. All prior art floating bridge driver topologies go against this teaching by operating the driver gain devices in such a way that they experience either large voltage swings or large current variations, or both. While U.S. Pat. No. 4,107,619 prescribes remedies such as cascoding and current bootstrapping for improving prior art circuits in this regard, these techniques can significantly reduce operating efficiency and increase circuit complexity, especially when used in a power amplifier.
b) Reduced output and operating efficiency. Even without cascading and current bootstrapping, inserting a driver stage in the signal path can reduce the maximum output voltage swing of the amplifier system by the amount of voltage needed to keep the driver gain devices in conduction. This can result in a loss of several volts of output swing, sufficient to reduce a 50 watt amplifier to 35 watts output, with a corresponding loss in operating efficiency.
c) Nonlinear loading of the preceding voltage amplifier output. Nonlinear behavior in a BJT or MOSFET driver gain device results in nonlinear impedance at the device input terminal, just as it does with the output gain devices. The addition of a driver generally reduces the magnitude of the problem, but does not eliminate it. This requires that the preceding voltage amplifier enforce linear operation by maintaining a much lower output impedance than otherwise would be needed, either intrinsically by operating at relatively high idle current, or through the use of negative feedback, which in practice requires that the voltage amplifier be located physically close to the current amplifier, where its radiated noise and heat could reduce the performance, and possibly the longevity, of the voltage amplifier.
Thus, it can be seen that adding a driver stage to a solid-state floating bridge amplifier under the prior art, while highly desirable, involves various compromises and trade-offs, potentially resulting in only limited performance improvement at the cost of increased complexity, reduced operating efficiency, and the need for close integration of amplifier circuits that might benefit from physical separation. The present invention overcomes these limitations and provides additional benefits.
Several objects and advantages of the present invention are:
a) To provide a floating bridge amplifier of improved open-loop performance but without significant increase in design complexity and without significant reduction in overall operating efficiency.
b) To provide an efficient, low-distortion floating bridge amplifier having almost rail-to-rail output voltage swing.
c) To provide a floating bridge amplifier having a high and linear driver stage input impedance, ensuring compatibility with a wide range of associated voltage amplifier circuits, and facilitating the physical separation of the voltage and current amplifiers where beneficial to the application.
Further objects and advantages of this invention will become apparent from the following descriptions of the preferred embodiments.
An audio amplifier of the present invention includes a solid-state floating bridge output stage and a preceding solid-state driver stage, both stages drawing their operating power from the same pair of floating DC power supplies, but through separate and independent signal-current paths whereby each driver gain device operates with minimal inter-electrode voltage swings and current variations. The output and driver stages, together with their common floating power supplies and bias source(s), form a current amplifier having low open-loop distortion, almost rail-to-rail output voltage swing, and high and linear input impedance. This current amplifier may be combined with a lowpower, high quality voltage amplifier, which may be separately enclosed to allow for remote operation or isolation from the radiated heat and electrical noise of the associated current amplifier. This improved topology enhances amplifier performance and utility without undue added design complexity or reduced operating efficiency.